Despite researchers' increasing acknowledgement that multiple pathways underlie the development of adolescent substance use, few studies have examined demarcating factors beyond antisocial behavior to define alternate pathways of risk for adolescent substance use. Notable within theories of "self-medication" and among targets of prevention programs, negative affect may define a second pathway of risk through which adolescents come to abuse substances as a means of regulating negative emotions. Although previous findings provide inconsistent support for the relation between negative affect and substance use among adolescents, these studies share three primary limitations that will be overcome by the current proposal. A FIRST NIH RESUBMISSION BY A YOUNG INVESTIGATOR, the current proposal includes three aims that examine (1) individual patterns of covariation in daily measures of negative affect and adolescent substance use, (2) whether adolescents are more likely to respond to negative affect by using substances within certain peer contexts (i. e., low friendship support), and (3) whether the relation between negative affect and substance use strengthens as adolescents experience the stress of the transition to high school and whether this relation is moderated by friendship support, coping efficacy and gender. A two-stage research design will achieve these aims through a school-based survey of 528 eighth graders in a single school district and a field-based study that follows 100 "at risk" youth as they transition to high school. Teens who have initiated substance use prior to leaving eighth grade and their best friends will be selected from the school based study to form the field-based sample. The field-based study will examine the feasibility of experience sampling and observational methods to examine whether daily covariation in patterns of mood and substance use emerge following the transition to high school, especially among those youth in less supportive friendships. Proposed statistical analyses include categorical random coefficient growth modeling to best capture questions of individual differences in change over time. Additional strengths of the proposal include assessment of an ethnically diverse sample of rural youth across the school year and, perhaps most uniquely and rarely targeted in studies of adolescent substance use, summer months. Findings will serve to determine the feasibility of these methods to inform subsequent studies of adolescent substance use.